Sovereign by Default
What it actually takes to build a runtime that answers to its user and not to a market.
“Sovereign” is a word that gets used loosely in software marketing, usually to mean “we are not evil.” It should mean something much more specific and much more demanding.
The actual definition
A system is sovereign when the user holds the keys and can verify what is running. Not “you can export your data.” Actually:
- You can run the whole thing without a network call to us.
- The binary you execute is inspectable.
- The system degrades to local-first, not to a modal error.
- Nothing in the runtime is optimized against your attention.
That last one is the hard part, and the only one that is genuinely scarce.
What this costs
Sovereignty is expensive in ways that do not show up on a feature grid:
- No telemetry you cannot personally switch off.
- No dependency you cannot vendor and audit.
- No server component that becomes load-bearing for correctness.
Each of those is a real engineering constraint, and each of them removes a revenue surface. This is why very few things are genuinely sovereign, and why the ones that are tend to be small.
The open-core line
The line I have drawn for LOGOS: the kernel, the calculus, the type system, and the runtime are open. The hosted coordination fabric is not.
The kernel is where sovereignty actually lives. If the kernel is open, then even a user who never touches our infrastructure can run their own instance and get the full experience. Everything above the kernel is convenience, and convenience should be optional.
If we ever made the kernel proprietary, the project would still be useful software. It would stop being sovereign software. Those are not the same thing, and the difference is the whole point.